INDEX

Explanations

words related to law enforcement or cop-related terms

oai_token-act-pair · gpt-3.5-turbo

references to police or law enforcement

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

sbm

-0.94

FORE

-0.93

Downloadha

-0.76

Accessory

-0.69

WAY

-0.69

Reloaded

-0.65

WAYS

-0.64

veyard

-0.64

inventory

-0.63

FINEST

-0.63

POSITIVE LOGITS

yrights

1.33

yright

1.10

rodu

1.09

enhagen

0.98

ious

0.97

yp

0.92

ulatory

0.82

eland

0.80

rol

0.80

resent

0.80

Activations Density 0.007%

words related to law enforcement or cop-related terms

references to police or law enforcement

No Comments

No Known Activations